Flexibility Platform Strategy — A Playbook for New Entrants

A strategic synthesis for a small, AI-accelerated team (≈10 people) building a flexibility platform service in the Swedish market. It assembles the empirical and regulatory picture across the wiki into a coherent answer to four questions: what to build (tech roadmap), how to get paid (business model), how to align with the coming rules (regulatory alignment), and how to win share against SWITCH and NODES. It closes with non-SaaS distribution models, which materially change what a 10-person team can do.

The Swedish market has two paradoxes that frame the whole exercise:

1. The market is thin but the addressable problem is growing. In winter 2025/26, 6 of 11 E.ON SWITCH markets had zero clearing; the entire Swedish flex market activated ~292 MWh over a full winter (Source - SWITCH Marknadsdata (info.switchmarket.se, 2026)). At the same time, DSO-aggregate flex needs grow 3–5× over the decade (277→2,523 MW; Source - Ei PM2025-03 DNDP Sammanställning (2025)) and E.ON alone reports ~700 MW / ~1 TWh of stated need and 17 GW of batteries in connection queue.
2. The two incumbents haven’t actually won. SWITCH is captive to E.ON Energidistribution — an Art. 32 neutrality problem the NC DR will sharpen. NODES is a neutral third party but has one live Swedish DSO (Effekthandel Väst), two discontinued ones (sthlmflex, JämtFlex), and — based on public evidence — a small Swedish team. Neither has locked the market, and the regulatory window 2026–2028 reshuffles the deck.

This is a market where the right 10-person team, launching in 2026, has a real shot. The rest of this page argues specifically where and how.

The strategic window: 2026–2028

Four regulatory events in the next 30 months remake the playing field:

The practical implication: a product built in 2026 to be NC DR-native and FIS-native from day one arrives just as incumbents face mandatory rework. SWITCH was designed pre-NC DR (2019 origin in CoordiNet) and carries architectural legacy — proprietary-API to OpenADR migration took until Q2 2024, CIM mapping is still a draft. NODES has never been publicly validated against NC DR Art. 38 and has no published roadmap (NODES › Data gaps).

A new entrant that is slower-footed than the next twelve months will miss the window. Incumbent rework will close it.

The opportunity map

The wiki surfaces seven distinguishable platform niches. Most analyses treat “the flexibility platform” as one thing; it is not. They have different customers, different unit economics, different technical cores.

The single highest-leverage niche for a 10-person team is #2 (FSP-side) combined with #4 (FIS layer) — see Three viable archetypes below. Neither is held. Both scale without requiring enterprise DSO sales motion.

The lowest-leverage play: #1 head-on competition. That requires replacing SWITCH at E.ON (politically impossible) or NODES at Göteborg (one contract, sticky, low revenue upside given Effekthandel Väst‘s ~930 MWh / 27 FSPs scale; Source - Göteborg Energi Elektrifieringsrapporten nr 1 (2025)).

Tech roadmap for a 10-person AI-accelerated team

The tech stack for a competitive flex platform is surprisingly small — the state of the art is visible in SWITCH and NODES, both of which are buildable by a ~10-person team today. SWITCH v1 went from concept to launched product in approximately 7 months; the years that followed were product iterations driven by changing market design, not platform complexity. AI acceleration (code generation, integration test synthesis, regulatory document parsing, compliance mapping) should compress even that timeline — a 2026 team building with current tooling can reach v1 feature parity well inside a year.

Build order — 18-month horizon

Months 0–3: Core primitives. Bid intake, merit-order clearing, impact-factor (PTDF) weighting, time-window configuration, settlement export. The mechanics are simple once decomposed. Target feature parity with SWITCH‘s market tool at ~40% of the code footprint by aggressively ditching SWITCH’s legacy.

Months 2–4: OpenADR 3.0.1 VTN + 15-min delivery. The 2025 industry supplement specifies two Swedish production variants (E.ON CONSUMPTION_POWER_LIMIT PT15M; Ellevio SIMPLE Curtail/Restore PT20M). Build VTN that supports both variants from the first release — this is the interoperability wedge against E.ON’s and Ellevio‘s asymmetric VTNs. See OpenADR. Note: this OpenADR layer covers Villkorade Avtal communication only (equivalent to ~2 months of SWITCH development). The broader market API (bid submission, metering, settlement) will remain proprietary REST until EU/Sweden agree joint data standards for flexibility markets — which had not happened as of 2026. Build your market API to be clean and documented; do not wait for a standard that may not arrive before your v1.

Months 3–6: Standardized product library. Implement LFM-h / LFM-p / LFM-e as a library, not hard-coded: A/B pricing models, dispatch obligation chain, upward-only 0.1 MW minimum, two-auction LFM-e (D-1 09:30 / H-2). Adopt them as the default product family — Ei approved them December 2025 and they are the product surface every DSO now standardizes on. Map SWITCH ST/TO/DO and NODES LongFlex/ShortFlex/MaxUsage to this canonical form.

Months 4–8: FIS-native data layer. Build CIM (IEC 62325) message types and API surface per NC DR Art. 24–28 from the start. Even before Ei+Svk publish the Swedish FIS spec in September 2026, the direction is known: flexibility register, CU/SP/SPU/SPG pathway, 4-year interoperability deadline. A platform that is FIS-native from release has permanent interoperability moat when the deadline bites incumbents.

Months 6–10: DSO grid-state integration. SWITCH requires DSOs to push actuals + forecasts + limits via API (Source - SWITCH DSO Onboarding Guide (user.switchmarket.se, 2026)). Building a forecast service here is harder than it looks: the historical and real-time power measurements needed for substation-level load forecasting are component-specific, DSO-owned, and treated as critical infrastructure — there is no public dataset suitable for this purpose (the Ei karttjänst has aggregate planning data, not operational measurements). DSOs treat this data as sensitive due to foreign-actor attack risk. The consequence is that most DSOs prefer building forecasts in-house, and third-party providers can only enter through a high-trust relationship with each DSO individually.

The practical implication for a new entrant: forecast-as-a-service is not a generic SaaS wedge. It is viable only as a value-added service bundled into a broader DSO relationship — after you already have the market platform deployed with that DSO. Build the capability (TFT or successor, grid-topology-aware); position it as an optional add-on that requires a data-sharing agreement, not a self-serve product. FlexAbility DR5 confirms small DSOs lack this capability — but the route to them is through trust, not a signup page.

Months 8–12: FSP-side multi-market bidding. Batteries in Effekthandel Väst have received “roughly equal revenue from FCR-D and local flex” (Source - FlexAbility Delrapport 2 (2025)). Value-stacking is unbuilt. Deliver:

• Unified view across SWITCH + NODES + FCR-D + mFRR + Nord Pool day-ahead
• BRP-aware bid scheduling (the D-2 10:30 shift SWITCH made in 2025/26 was forced by aggregator BRP workflow)
• Portfolio optimization across resources
• Baseline integrity auto-detection (learn from the V2022/23 SWITCH baseline-tampering incident)

Months 10–14: Cybersecurity / beredskapsflexibilitet layer. Energimyndigheten ER 2025:35 explicitly flags synchronized heat-pump activation as systemic risk (Nordic32 simulation, ~300k connected heat pumps at critical mass). Build in random startup delay (UK 600s mandate as reference), rate limiting, and activation jitter as platform-level primitives — this is the Art. 24 Implementing Act’s second-pillar territory and a compliance-driven moat.

Months 12–18: TSO integration. Svenska kraftnät SUSIE bridge for subscription requests (SWITCH has this; it is valuable but straightforward). BSP-aware settlement for when the BSP role is finally implemented (DR5: BSP deferred to 2028, +300 MW realizable on day one — worth positioning for early.

Do not build mFRR forwarding. Both SWITCH and NODES attempted this during CoordiNet: SWITCH via API-to-Svk-SFTP, NODES via equivalent integration. Both scrapped it. Svk provided minimal support, timelines stretched, and the implementation was sub-par. The technical path exists; the institutional support does not. Revisit only if Svk formally opens an integration program.

10-person team shape

The pattern: AI acceleration is highest where the work is structured (code, SDKs, regulation mapping, test synthesis). It is lowest in DSO relationships, FSP onboarding, and policy participation — which is precisely where FlexAbility DR5 interviews indicate real bottlenecks live. A small team’s biggest leverage comes from eliminating the work where AI is best and concentrating human time on the relationships.

Business model: infrastructure pricing, not transactions

The most common strategic error for a flex platform: betting on transaction fees. The Swedish data kills that model.

The thin-market math. V2025/26 activation across all markets was ~292 MWh (Source - SWITCH Marknadsdata (info.switchmarket.se, 2026)). Even at 5% transaction fee on 5,000 SEK/MWh average activation price, the entire Swedish market generates ~73,000 SEK/year. Add availability compensation (~10 MSEK across all markets) and a 2–3% fee: ~200–300 kSEK. The entire Swedish flex market at transaction fees cannot fund one engineer. This is why NODES is small and SWITCH is an E.ON cost center.

The correct unit economics are infrastructure pricing — recurring fees per DSO, per FSP, per grid-point, per qualification — decoupled from transaction volume. Each pricing anchor serves a different strategic purpose.

Outcome-based pricing is the interesting long-game bet. FlexAbility DR3 quantifies DSO value cases: 10,711 SEK/MWh utnyttjandegrad, 122,000 SEK/MWh for first-MW abonnemangsoptimering, 1.09 MSEK/MW average grid-upgrade cost. A platform that contracts on “X% of verified deferred CAPEX” aligns perfectly with TOTEX lösningsneutralitet from 2028. The barrier is that pre-2028 CAPEX bias makes DSOs averse to strategies that reduce their rate base. Post-2028 this inverts. Build the measurement infrastructure in 2026–2027 (forecast counterfactuals, CAPEX deferral quantification) and you own the commercial model when the regulation flips.

Rule of thumb for a 10-person team: target 2–3 MSEK/year/large-DSO infrastructure fee for the full stack (market + FIS + villkorade avtal VTN + forecasting). Five large DSO contracts = 10–15 MSEK ARR, which funds the team at Swedish salary levels without venture capital. Ten contracts including middle-tier DSOs = profitability.

Regulatory alignment: treat it as the primary product

The standard startup playbook treats regulation as a constraint; in this market, regulation creates the demand. FlexAbility DR5 ranks intäktsreglering as the #1 barrier at 38 points — above all others combined. DSOs describe flex procurement as “välgörenhet” (charity). The path from that world to a healthy flex market runs entirely through regulation.

Specific alignments to build around. All are sourced to Source - NC DR Amended Text (ACER Recommendation 01-2025 Annex 1) unless noted:

• Art. 24–28 FIS. Four-year interoperability deadline. Build CIM-native data model from day one. This is the single highest-leverage regulatory choice: whatever Ei+Svk land on in their September 2026 report, a FIS-native platform slots into it.
• Art. 29 / 30 / 31. Market-based default + derogation framework + flexible CAs must coordinate with market procurement. Build the platform to treat Villkorade Avtal and market bids as a unified resource pool with rules about which pathway to call first. Ei’s 2025 ställningstagande already says villkorade avtal are a non-market Art. 13(3) tool: the platform that can audit the DSO’s choice between market-first and villkorade-avtal-first on a per-event basis wins Ei’s compliance-reporting demand.
• Art. 37 transparency. 1-day publication window, single national access point. Build public data publishing as default; adopt info.switchmarket.se-style openness as the baseline. This costs nothing and preempts a compliance mandate.
• Art. 38 product harmonization. 14 mandatory attributes. Since LFM-h/p/e already map cleanly, there’s no migration cost for a greenfield platform. For SWITCH and NODES, whose products predate the standard, retrofitting is real engineering work.
• Art. 43–44 DNDP integration. Platforms need to export flex-needs data in the form DSOs report in their DNDPs and in FNA Tabell 15. Build the Ei karttjänst format as a native export — this makes the platform useful for a DSO’s DNDP cycle, not just its operational season.
• Art. 24 Implementing Act (parallel track). Energimyndigheten ER 2025:35 surfaces this as the second pillar alongside NC DR. Cybersecurity + protocols + random startup delay + beredskapsflexibilitet. A platform that is “ready for the Art. 24 IA” differentiates against SWITCH/NODES which were designed pre-cybersecurity-framing.

Ei T&C participation is non-optional. Ei’s NC DR preparation has Yuri Joelsson designing terms and conditions from November 2025. The team that is in the room with Ei during 2026 T&C drafting shapes what “compliant” means. This is a ~10% time commitment for the regulatory/product lead and disproportionate strategic leverage. Energiföretagen‘s AG Helhet Flex is the parallel industry forum — equally important.

Taking market share: where to wedge in

Do not attack SWITCH in Skåne. Södra Skåne is E.ON’s dominant revenue market (377,600 SEK/MW seasonal rate, 60% above all other markets). E.ON will defend it with captive-DSO advantage, ectocloud district heating integration, DIS procurement lock-in, and impact-factor sophistication. It’s a fortress.

Do not attack NODES in Göteborg. Effekthandel Väst is NODES’s single active Swedish contract. Göteborg Energi Nät has operational familiarity, Season 4 data (930 MWh, 27 FSPs, world-first V2G), and a stable relationship. Replacing NODES there requires unambiguous superiority at ~5 MSEK switching cost — unavailable to a new entrant.

Four real wedges:

Wedge 1: Long-tail DSOs (135+ small nätföretag)

Ei’s PM2025:03 covers 155 DSOs. The four large DSOs (E.ON, Vattenfall, Ellevio, Göteborg Energi) represent ~70% of customers. The remaining ~150 small/medium DSOs represent 30% of customers and — based on PM2025:03 — largely no flex capability at all.

NC DR forces these DSOs into the market by 2028. They cannot afford a SWITCH-scale bespoke platform. Their volume per market is too low for NODES’s SaaS pricing to make sense. They need cheap, turnkey, compliance-grade infrastructure.

Product: a standardized LFM-h/p/e market-in-a-box with Ei-approved products pre-implemented, forecast-as-a-service, VTN for villkorade avtal, FIS export — priced at ~200–500 kSEK/year per small DSO. Fifty small DSOs at that price = 10–25 MSEK ARR. This is the single largest greenfield segment.

Wedge 2: FSP-side multi-market optimization

Aggregators, battery operators, and large industrial FSPs currently have no integrated tooling across SWITCH, NODES, FCR-D, mFRR, Nord Pool, and (soon) the national BSP role. The Swedish evidence is clear: FSPs need this (CoordiNet showed FSPs double-registered to capture both mFRR and local flex), the platforms don’t provide it (SWITCH focuses DSO-side; NODES same), and the revenue is real (FlexAbility DR2 shows batteries earning ~equal revenue from Effekthandel Väst and FCR-D in 2024/25).

A platform that lets an FSP manage a 10 MW portfolio across 5 markets simultaneously, with BRP-correct timing and settlement consolidation, captures meaningful per-FSP revenue (20–50 kSEK/year per FSP, 200+ FSPs addressable). More importantly, it owns the supply side — which gives leverage in every DSO negotiation that follows.

Data access prerequisite. An FSP OS platform serving customers across multiple DSO territories needs to register as Berättigad part with each DSO to access 15-minute meter data independently — standardized via PRODAT Z13–Z18 since June 2025, replacing the prior case-by-case power-of-attorney process. The bilateral agreement overhead (up to ~170 DSOs) is a real but budgetable cost of entry. Build berättigad part registration into the market-entry checklist; the per-DSO fee and agreement process are known quantities. Until DHV/FIS consolidates data access centrally (post-2028), this is the operational path. (Berättigad part)

Wedge 3: Third-party forecast and grid-state provider (conditional)

SWITCH’s documentation names third-party forecast providers as an integration layer (Source - SWITCH DSO Onboarding Guide (user.switchmarket.se, 2026)). E.ON built TFT in-house; smaller DSOs cannot. In principle this is a real wedge.

In practice the data access problem is severe: the historical and real-time substation measurements needed for short-term load forecasting are DSO-owned critical infrastructure, not publicly available. No dataset exists outside each DSO’s own systems. DSOs guard this data against foreign-actor attack risk and strongly prefer keeping forecasting in-house. Third-party access requires a bilateral data-sharing agreement negotiated at high trust level.

This means forecast-as-a-service is not a standalone SaaS wedge. It is viable as a value-add bundled inside a broader market platform relationship: once a DSO has deployed your platform and a data-sharing agreement is already in place, adding managed forecasting is natural. As a cold approach to a DSO that uses SWITCH or NODES, it is a very long sales cycle with no obvious entry point. Pursue it as upsell, not as wedge.

Wedge 4: Production-side and V2G (not a real wedge)

SWITCH already supports production-side markets — Projekt Halland (summer 2025) uses the same platform with downregulation mode enabled. There is no product gap here for a new entrant to fill.

The demand side is also much weaker than for winter consumption markets: Sweden’s transmission grid still has significant margin in summer, making production-side congestion a localized and seasonal problem rather than a structural one (Ellevio’s Skaraborg 0–450 MW figure is a peak-conditional 10-year estimate, not a near-term operational need). V2G demonstrated its first Swedish delivery in March 2025 (Source - Göteborg Energi Elektrifieringsrapporten nr 1 (2025)) but remains tiny in volume.

Do not position a new platform primarily around production-side or V2G. These are features to support; they are not market entry points.

Alternatives to SaaS for distribution

Pure SaaS has three problems for a 10-person team in this market:

1. Sales motion is enterprise-long. DSO procurement (LOU/LUF-constrained) is 12–24 months per contract. A 10-person team cannot run five parallel enterprise sales cycles.
2. CAC is high and payback is slow if pricing is infrastructure-level (2–3 MSEK/year, not 50 MSEK).
3. The customer’s neutrality concerns scale with the platform company’s size, especially under NC DR Art. 32. A small startup hosting market clearing for 10 DSOs raises regulator eyebrows.

Non-SaaS distribution models address these directly. Each has a strategic logic; they are not mutually exclusive.

1. Open-core

Release the market-tool primitives, OpenADR VTN, and LFM-h/p/e library as permissive open source (MIT/Apache). Charge for the commercial extensions: forecast-as-service, FIS bridge, managed hosting, compliance audits, SLA support.

Strategic logic: distribution becomes free. Any DSO, integrator, or consultant can start with the OSS core; upgrade to commercial when they want support or advanced features. A single OSS implementation at Jämtkraft (see JämtFlex) or a small Luleå Energi-class DSO becomes a reference installation — the same code base that serves the commercial customer.

Swedish precedent: none in this domain; broadly applicable elsewhere (GitLab, Elastic, Redis before the relicensing era). OpenADR itself is an open standard; building on it as OSS amplifies ecosystem effects.

Risk: forks. Mitigation: keep the commercial differentiator (FIS, forecast ML, compliance) under commercial license.

2. DSO cooperative / mutual ownership

Several DSOs co-own the platform entity. Each pays an annual subscription proportional to customers; decisions are consensus at a platform-governance board. The platform has no external shareholders.

Strategic logic: directly addresses Art. 32 neutrality concerns (no single DSO controls; no commercial incentive to favor a DSO). Sweco’s 2025 mapping implicitly pointed this direction with its concern about fragmented non-neutral infrastructure.

Swedish precedent: Elhubben governance (originally proposed as Svk-owned; now datahanteringsverktyg to Ei + Svk); Ellagret (Energiföretagen-owned industry tooling); Nord Pool’s original cooperative history.

Operating model: the 10-person team is the operating company, compensated on a long-term contract (5–10 years) with the cooperative. The team is not the owner; it is the operator. This removes capital pressure and converts the business model to long-horizon stable.

Who sponsors: the 20% of DSOs (PM2025:03) already using flex services could mutualize infrastructure costs that each currently bears individually.

3. Regulated shared utility (FIS-native)

The datahanteringsverktyg assignment is already building the scaffolding: Ei + Svk will report by 30 September 2026 on the scope of a centralized data tool with FIS as a named use case. A team that contributes to that scope design — technical prototypes, architecture papers, sandbox code — can position to be the operating vendor of the resulting FIS.

Strategic logic: Sweden gets one FIS. If it is regulated infrastructure, the operating vendor is chosen by tender or cost-recovery model, not commercial sales. The winner operates at regulated margin but with guaranteed volume and no CAC. Elmarknadshubb went this way before being paused.

Distribution logic: no distribution problem. The regulator mandates it; every DSO and FSP connects.

Risk: procurement timelines are slow; there is no guarantee a new entrant wins. Mitigation: engage early (2026) in the scope-setting phase; become the reference implementation the regulator trusts.

4. White-label / embedded library

Ship the platform as a library that DSOs (or their existing IT vendors: Tietoevry, Trimble, Nodes of the non-flex kind) embed inside their own operations stack. The library exposes APIs; the DSO-facing UI is built by the embedder.

Strategic logic: converts a DSO relationship problem into a vendor relationship. E.ON explicitly licenses SWITCH to “andra aktörer som är i behov av ett verktyg” (Source - E.ON Nätutvecklingsplan 2025-2034). A white-label offering competes on neutrality and cost per DSO seat.

Pricing: per-developer or per-deployed-instance license fees, plus support contract. Enterprise-software pricing.

5. Appliance / on-premises / sovereign cloud

Deliver the platform as a containerized deployment the DSO hosts on its own infrastructure or in Ei-sovereign Swedish cloud. No data leaves the DSO’s environment. Monthly or annual license fee.

Strategic logic: cybersecurity-grade trust. Energimyndigheten ER 2025:35 flags cybersecurity as a central concern; SÄPO and FRA reviewed the datahanteringsverktyg scope (Source - Uppdrag Centralt Datahanteringsverktyg (2025)). A DSO that cannot legally put load forecasts or FSP bids in US hyperscaler cloud needs this option.

Pricing: 500 kSEK–2 MSEK/year per deployment, depending on scale; higher than SaaS per-DSO but lower than building in-house.

6. Industry consortium spin-out

Energiföretagen, via its AG Helhet Flex working group, already coordinates DSOs on flex matters. A consortium commissions the platform build; the consortium owns the output; the operating team is a contractor.

Strategic logic: Energiföretagen has the trust, audit credibility, and standards-setting role (LFM-h/p/e product design). A platform blessed by Energiföretagen comes with effective pre-sale at every member DSO.

Operating model: similar to cooperative (option 2) but with Energiföretagen as the governance hub.

7. Outcome-based / shared-savings

Price on verified CAPEX deferral: the platform is paid X% of the investment avoided, measured against a defined counterfactual. Most radical; requires the platform to host or access CAPEX plans, build counterfactual modelling, and negotiate recognition of flex as an Art. 32 alternative.

Strategic logic: aligns perfectly with RP5 TOTEX lösningsneutralitet (2028). Pre-2028 it is a hard sell because CAPEX bias rewards the DSO’s investment directly; post-2028 it becomes the natural pricing model.

Preparation: build the counterfactual measurement infrastructure in 2026–2027 (the BeFlexible D5.2 congestion forecast evaluation tool is an instructive reference).

8. Data intermediary / broker model

The platform does not operate markets directly. It aggregates market, grid, and resource data — from Ei karttjänst, SWITCH public portal, ENTSO-E, NODES (public where available), DSO APIs — and sells cleaned, analyzed data to FSPs, regulators, aggregators, investors, and consultants.

Strategic logic: completely platform-agnostic; benefits from market fragmentation rather than competing with it. Low CAC (subscription, typed through a simple web portal). Becomes the Bloomberg terminal of Swedish flex.

Pricing: 100–500 kSEK/year per subscriber. Realistic 50–100 subscribers (aggregators, banks, consultants, regulators). Total addressable: 10–50 MSEK/year.

AI acceleration: extreme. Data extraction, cleaning, analysis, summarization, trend detection — all structured work. A 3-person data team with LLM infrastructure can serve this market.

Three viable archetypes

Synthesizing the above into three coherent, internally consistent ten-person plays. These are not shopping lists — each requires commitment to one strategic posture.

A. The Neutral Cooperative Utility

Premise: build the long-tail DSO platform as a DSO-owned cooperative under Energiföretagen sponsorship; bundle LFM-h/p/e market tool + OpenADR VTN + forecast-as-a-service + FIS bridge; operate at regulated-utility margins.

Distribution: no SaaS sales. Energiföretagen AG Helhet Flex sponsors; 5–10 small/medium DSOs co-found; the operator is the 10-person team on 10-year contract.

Product: LFM-h/p/e native; FIS-native; CIM-native; neutral-by-governance.

Revenue horizon: 5–15 MSEK ARR by year 3; slow, stable, high retention.

Wins against: SWITCH (neutrality); NODES (Swedish ownership, lower per-DSO cost).

B. The FSP Operating System

Premise: serve aggregators, battery operators, and large industrial FSPs with a single operating layer across SWITCH + NODES + FCR-D + mFRR + Nord Pool + (future) BSP. Portfolio optimization, BRP-aware scheduling, baseline integrity, settlement consolidation, regulatory reporting.

Distribution: pure SaaS with self-serve signup; low-friction onboarding targeting the ~20–50 active Swedish FSPs and the next wave of new FSPs entering via the BSP role and V2G in 2027–2028.

Product: SDK in TypeScript + Python; portal + API; LLM-assisted market intelligence; compliance reports.

Revenue horizon: 10–30 MSEK ARR by year 3 if Swedish FSP count doubles; substantial upside if product works in Norway, Denmark, Finland, Germany.

Wins against: nothing — this niche is unheld.

C. The Flex Data Broker

Premise: become the neutral data intermediary for Swedish flex markets. Data ingest from every public source (Ei karttjänst, SWITCH portal, ENTSO-E, DNDP filings, FNA publications, Svk SUSIE stats) plus contracted access to DSO data where available. Sell processed data + analysis + regulatory intelligence to FSPs, regulators, banks, consultants, and investors.

Distribution: low-touch SaaS subscription; content marketing via a public blog that becomes the reference commentary on Swedish flex markets (SWITCH, NODES, RP5, NC DR, FNA).

Product: data product (feed, API, dashboards) + research product (reports, LLM agents for ad-hoc analysis).

Revenue horizon: 3–10 MSEK ARR by year 3; low operational complexity; high per-employee revenue; strong AI acceleration fit.

Wins against: nothing — this niche is also unheld.

Which archetype for which founder

Archetype B has the cleanest technical win and the biggest outside-Sweden upside but requires direct FSP relationships. Archetype A has the deepest moat but requires political capital with Energiföretagen and Ei. Archetype C has the lowest barrier to entry but also the smallest addressable market — best as a founding-capital step toward A or B.

The case for combining B + C: the Data Broker buys the team time and analytical reputation while the FSP Operating System is being built. The data product and the FSP product share a data pipeline. This hybrid is what a 10-person AI-accelerated team can plausibly execute in 2026–2027.

What to avoid

• Chasing Södra Skåne / Effekthandel Väst-style thick markets. They are incumbent fortresses and their apparent size is misleading — total Swedish flex activation is tiny, and the revenue concentration at the top masks thin-market pathology.
• Transaction-fee pricing. The 292 MWh total winter activation makes this arithmetically impossible as a standalone model.
• Reinventing OpenADR, CIM, or LFM-h/p/e. These are fixed costs the ecosystem has already paid. Adopt them; differentiate elsewhere.
• Building before the FIS specification is clear (waiting out September 2026). Waiting is also wrong — the direction is known from NC DR Art. 24–28. Build on the known 80% and adapt the 20%.
• Treating E.ON as the customer. E.ON has SWITCH; it is the competitor, not the buyer. The 150 small/medium DSOs and 20–50 active FSPs are the buyers.
• Over-indexing on pure SaaS. The cooperative, white-label, and outcome-based models are better fits for the small-team + regulated-market combination.

The underlying strategic insight

The ten-person team’s advantage in this market is not speed, feature count, or AI productivity in the abstract. It is the ability to arrive with the right regulatory architecture in place the year the regulation requires it — while incumbents retrofit legacy systems into a framework the regulation didn’t anticipate.

SWITCH won the 2019–2025 period by being first and internal-to-E.ON. NODES won the neutral-platform slot by default. Neither won the 2026–2030 period. That period is up for grabs, and it belongs to the platform that is NC DR-native, FIS-native, and LFM-native from day one, with the patience to build infrastructure-priced revenue slowly while thin markets mature. A ten-person team with AI acceleration can build exactly that.

Related pages

• SWITCH — incumbent, E.ON-captive, architectural legacy
• NODES — incumbent, Norwegian, Swedish footprint limited to Effekthandel Väst
• Flexibility Market — market landscape and empirical outcomes
• Network Code on Demand Response — the regulatory trigger for 2026–2028
• Elmarknadshubb — the FIS vehicle in development
• OpenADR — the protocol incumbents inherit; a greenfield platform builds on 3.0.1 from day one
• Flexibility Communication Protocols — the full stack; four layers; Art. 24 Implementing Act
• Why Swedish Local Flex Markets Are Thin — Structural Causes — why infrastructure pricing is forced
• DSO Flexibility Valuation — Methods and Swedish Evidence — the DSO willingness-to-pay that outcome-based pricing captures
• NC DR Implementation in Sweden — What Changes and for Whom — the actor-level regulatory map

Data gaps

• NODES AS revenue, team size, and ownership structure (commercial terms not public)
• SWITCH licensing commercial terms — whom has E.ON licensed to, at what price
• Whether Ei+Svk datahanteringsverktyg scope will permit third-party FIS operators or mandate a regulated utility
• Norwegian centralized flex register outcome — does it validate the cooperative model
• Realistic addressable FSP count in Sweden by 2028 (baseline 20–50 today; aggregation + BSP + V2G could 3–5× this)
• Pricing benchmarks for forecast-as-a-service in Swedish grid context
• Whether any of the 155 DSOs have explored cooperative platform procurement as an alternative to SaaS
• RP5 TOTEX lösningsneutralitet numerical design — how the benchmarking actually scores flex vs CAPEX