Fermi Inc. (NASDAQ & LSE: FRMI) is having the kind of month that separates “big idea” infrastructure stories from the cold reality of signed contracts and cash timing.
The headline catalyst: a prospective first tenant for Fermi’s flagship Texas development, Project Matador, terminated an agreement that could have advanced up to $150 million toward construction costs—triggering a violent market reaction and a fresh round of debate about what FRMI is worth before meaningful revenue arrives. [1]
Below is what’s known right now (as of December 15, 2025), what analysts are forecasting, and what to watch next.
What happened to FRMI stock — and why it mattered so much
On December 11, 2025, Fermi said the “First Tenant” notified the company that it was terminating the AICA—an Advance in Aid of Construction Agreement tied to leasing part of the Project Matador site. Fermi disclosed that no funds had been drawn under the agreement before termination, and that the parties were still negotiating lease terms under the earlier letter of intent. [2]
Markets did not applaud the nuance.
On Friday, December 12, FRMI shares suffered a steep selloff. The stock closed at $10.09, down 33.84% on the day, after trading as low as $8.30 intraday, with volume surging to roughly 63.1 million shares. [3]
Reuters summarized the core issue plainly: a newly listed data-center real estate/infrastructure company just lost (at least temporarily) a major piece of early-stage funding support from an undisclosed customer meant to help finance construction at its Texas site. [4]
The key timeline (confirmed)
- Sep. 19, 2025: Fermi enters a non-binding letter of intent with an “investment grade-rated” prospective tenant. [5]
- Nov. 3, 2025: Parties sign the AICA, under which the tenant could advance up to $150M (subject to conditions). [6]
- Midnight Dec. 9, 2025: Exclusivity period in the LOI expires. [7]
- Dec. 11, 2025: Tenant notifies Fermi it is terminating the AICA; Fermi says lease talks continue and it has begun discussions with other potential tenants for 2026 power delivery. [8]
Why the market reacted so harshly: FRMI’s story is “contract-first”
Fermi is not being valued like a mature REIT with stabilized rent rolls. It’s being valued like a high-voltage promise: build a behind-the-meter “private grid” campus big enough to feed hyperscale AI compute, then lock in long-duration tenants.
That model makes early counterparties disproportionately important. Losing a funding commitment (even if the tenant is still “in discussions”) is a direct hit to perceived execution certainty—especially because Fermi was pitched to many investors as a front-row seat to the AI power crunch. [9]
The Financial Times reported that analysts tied the breakdown to pricing negotiations, suggesting Fermi refused to compromise on pricing to avoid weakening future deals—an argument that might make long-run sense, but can still hurt in the short run if it delays the first definitive lease. [10]
What Fermi says is still on track
Fermi’s public messaging has been consistent on two points after the AICA termination:
- No AICA money had been drawn (so it’s not a “cash clawback” event). [11]
- The company remains confident it can meet its expected Project Matador power delivery schedule, citing robust demand for behind-the-meter power for AI. [12]
Whether the market believes that confidence statement will depend on what comes next: a signed lease, a replacement tenant, project financing milestones, and continued evidence that power delivery infrastructure is progressing.
Other current Project Matador developments investors are watching
Even with the tenant funding shock, Fermi has continued to stack up “infrastructure progress” headlines—important because power availability (not just land) is the bottleneck for AI-scale data centers.
Xcel Energy agreement: up to 200 MW starting January 2026
Fermi announced a definitive Electric Service Agreement with Southwestern Public Service Company (SPS), a subsidiary of Xcel Energy, to provide up to 200 MW to the Project Matador campus, beginning with 86 MW in January 2026 and ramping over time. [13]
That matters because it’s one of the more concrete near-term “power-to-site” steps in a project that ultimately aims for gigawatts.
Cooling technology MOU (water use is political in West Texas)
Fermi also announced a non-binding MOU with MVM EGI Zrt. to develop hybrid cooling towers—positioned as reducing evaporative water use by 80%+ versus conventional all-wet cooling, with the first tower targeted for construction starting January 2026 and full system completion targeted by 2034. [14]
Cooling sounds mundane until you remember that, at AI scale, cooling becomes a permitting and community-acceptance issue as much as an engineering issue.
Context: what Fermi is building (and why the story attracts so much heat)
Fermi is developing Project Matador in the Texas Panhandle, aiming to integrate major power generation (including natural gas and nuclear, plus other sources) to supply AI and hyperscale computing customers. The company markets itself as building next-generation private electric grids at gigawatt scale for AI workloads. [15]
It’s also a politically loud story: Project Matador includes the Donald J. Trump Generating Plant branding, and the firm was co-founded by former U.S. Energy Secretary Rick Perry—facts that have amplified attention (and volatility). [16]
The state of FRMI forecasts: analysts are still bullish — but targets are being reset
Here’s the interesting tension in the FRMI tape right now:
- The stock is pricing in a major “execution haircut.”
- Many Wall Street analysts are still publishing Buy/Strong Buy views with very large upside from current levels.
That can happen when a stock drops faster than analysts revise models, but it can also reflect a genuine belief that tenant churn is normal early, and the bigger scarcity is reliable power.
Consensus targets (as of mid-December 2025)
Multiple tracking services show broadly similar averages:
- MarketBeat: average target $31.56, with a range of $27 (low) to $37 (high). [17]
- TipRanks: average target $31.63, with a $27–$37 range, and a Strong Buy consensus based on the analysts it tracks. [18]
- StockAnalysis: 12-month target $30.86, rating “Strong Buy” (per its analyst summary). [19]
- TradingView: published targets in the same general zone (around the low-$30s), depending on the snapshot and source feed. [20]
The big caveat: targets like these are highly sensitive to the probability of landing anchor tenants on attractive terms and the timeline to revenue. When the market starts doubting either, the “fair value” range can compress quickly.
Specific analyst moves now in the spotlight
- Evercore ISI: Investing.com reported Evercore cut its target to $20 from $37 while maintaining an Outperform rating after the AICA termination. [21]
- Stifel: a Futu/TipRanks-cited item reported Stifel maintained a Buy with a $29 target. [22]
- A compiled list of earlier initiation targets (late October 2025) circulated by market data services includes targets such as $27 (Mizuho/Cantor) up to $37 (Berenberg), with Macquarie cited at $35, among others. [23]
Meanwhile, commentary from mainstream market outlets has leaned “cautiously optimistic, but…” Barron’s emphasized that Fermi is early-stage and pre-revenue, and that the loss of the first major tenant funding commitment clouds near-term confidence even if the long-term AI-power theme remains strong. [24]
So… is FRMI a broken story or a delayed story?
Right now it’s best modeled as a probability tree, not a single narrative.
Bull case (what has to go right):
- Fermi signs a definitive lease with the original tenant or replaces it with another investment-grade tenant.
- Power delivery steps (like the Xcel/SPS ramp beginning Jan 2026) stay on schedule.
- Project financing progresses without heavily diluting shareholders.
- The broader AI buildout continues to pressure utilities and make behind-the-meter power more valuable.
Fermi itself points to continued tenant discussions and sustained demand for AI power as the reason it expects to keep timelines intact. [25]
Bear case (what can still go wrong):
- The “First Tenant” situation becomes a slow-motion unraveling (lease never finalized, other prospects wait on proof).
- Pricing power turns out weaker than expected once customers have real alternatives.
- Capex and financing needs force repeated capital raises (dilution) before revenue arrives.
- Permitting, grid interconnect, equipment delivery, or nuclear timelines slip—common failure modes for mega-infrastructure.
The Financial Times framed the latest setback as serious but not necessarily thesis-ending, noting that company advisers and some analysts still argue the long-term need for power-heavy AI infrastructure remains strong. [26]
Legal aftershocks: investor investigations appear, as they often do after sharp drops
After the December 12 collapse, multiple law firms published announcements about investigations into potential securities-law issues tied to disclosures around the tenant agreement(s). [27]
These notices are common following large one-day declines and are not proof of wrongdoing by themselves—but they can add headline risk and keep volatility elevated.
What to watch next (the practical checklist for Dec. 15 and beyond)
If you’re tracking Fermi Inc stock as a news-driven name, these are the next catalysts that matter more than day-to-day price noise:
- A definitive lease (with the First Tenant or a new anchor tenant)
The market wants a signed, bankable agreement—something that survives diligence, not just intent. - More detail on tenant economics
The big unanswered questions are pricing, term length, and what share of infrastructure costs tenants will bear. - Financing milestones
Fermi previously confirmed it was in discussions regarding $4+ billion in project financing for the first tenant (as reported in an earlier RNS statement). Any update here will move the stock. [28] - Operational execution into early 2026
The Xcel/SPS delivery schedule (86 MW starting January 2026) is a concrete near-term checkpoint. [29] - Index / flows narrative
Market chatter has pointed to FRMI joining major Russell indexes later in December (as reported by market news services). Inclusion can create mechanical demand, but it doesn’t fix fundamentals. [30]
Bottom line
As of December 15, 2025, Fermi (FRMI) is in a high-stakes “prove it” phase: the vision (AI power at scale) is still attractive, the execution story has real progress markers (like the Xcel/SPS power agreement), but the stock is now trading under the shadow of a very specific question:
Can Fermi turn gigawatts and acreage into signed tenants fast enough—and on terms strong enough—to justify the valuation?
References
1. markets.ft.com, 2. markets.ft.com, 3. stockanalysis.com, 4. www.reuters.com, 5. markets.ft.com, 6. markets.ft.com, 7. markets.ft.com, 8. markets.ft.com, 9. www.ft.com, 10. www.ft.com, 11. markets.ft.com, 12. markets.ft.com, 13. markets.ft.com, 14. www.prnewswire.com, 15. investor.fermiamerica.com, 16. www.marketwatch.com, 17. www.marketbeat.com, 18. www.tipranks.com, 19. stockanalysis.com, 20. www.tradingview.com, 21. www.investing.com, 22. news.futunn.com, 23. finviz.com, 24. www.barrons.com, 25. markets.ft.com, 26. www.ft.com, 27. www.businesswire.com, 28. markets.ft.com, 29. markets.ft.com, 30. seekingalpha.com
