Biotech catalyst, news and analysis PDUFA tracker
From narrative-rich AI photonics story to execution test: POET now has a far bigger balance sheet, a broader partnership map and a clearer 2026–2027 roadmap — but the market still wants the same proof as always: real shipments, repeat revenue and evidence that the Optical Interposer can scale outside press releases.
Applied Digital is one of the more interesting infrastructure stories in the market because it sits at the intersection of three forces that investors have spent the last year chasing aggressively: hyperscaler demand, power scarcity, and the belief that the economic winners of the AI cycle will not only be the model builders but also the owners of the physical capacity that makes those workloads possible. That is the seductive version of the story, and it is easy to understand why it caught such a bid.
Revenue fell 44% year over year to $10.3M, while non-GAAP EPS came in at a loss of $0.05. At the same time, quarterly bookings surged to $37.2M and effective backlog reached a record $50.9M. Management also said full-year FY26 revenue should land on the high side of the previously guided $45M-$50M range, while second-half bookings should come in at the high end of the prior $60M-$80M range.
POET arrived at the industry’s key optical-communications event with fresh news in hand: the new co-development announcement with LITEON, plus demonstrations of Blazar and Starlight products aimed at AI-related optical connectivity. In practical terms, this is the first immediate stage where the market can gauge whether the latest headline gets traction beyond the press release itself.
The market is no longer treating Middle East escalation as a passing geopolitical shock. It is increasingly pricing it as a broader economic regime: one that reprices oil and LNG, lifts defense and dual-use space names, pressures airlines and small caps, and forces investors to rethink the physical vulnerability of cloud infrastructure
Zeta Global ended 2025 with roughly $1.3 billion in revenue, high-20s growth, expanding margins and a still-modest GAAP net loss, while launching a strategic collaboration with OpenAI to power Athena, its enterprise AI marketing agent. The investment case now revolves less around “can they grow” and more around “can they sustain growth, expand margins and prove that their data and AI stack is truly differentiated.
A detailed look at Palladyne AI’s fiscal 2025 numbers, backlog and capital structure after the latest annual report and guidance update.
The company prints a sharp revenue decline but a much narrower net loss, closes Ask Sage and CargoSeer, and walks into 2026 with over $460 million in cash and investments — at the price of massive dilution and a complex capital structure. Retail is back in the name, but the story is anything but low-risk.
IREN Limited, formerly Iris Energy, is rapidly transforming itself from a cyclical Bitcoin miner into a vertically integrated AI infrastructure provider, combining long-duration power contracts, hyperscale data centres and multi-year GPU cloud deals with blue-chip counterparties. The next phase of the story is all about execution: how fast the company can switch its asset base to AI, how stable those cloud revenues become and how it balances growth ambitions with leverage and dilution.
BigBear.ai is trying to turn years of work in defense intelligence, mission planning and logistics optimization into a scalable AI platform story, boosted by large capital raises and the acquisition of generative-AI partner Ask Sage. Revenues are meaningful and the cash cushion is large, but the path from contract wins to durable, profitable growth remains complex and execution-dependent.
Updated overview of POET Technologies after the 2025 production orders, equity financings and the 2026 Lightwave award: technology, product line, financial profile, partnerships, retail sentiment and key risk factors.
SoundHound AI started as a voice and music-recognition company and gradually evolved into a broader conversational-AI provider. It now positions itself as a global platform for voice and agentic AI, powering restaurant phone ordering and drive-thrus, in-car voice assistants, smart devices and enterprise customer-service agents across financial services, healthcare, retail and more.
Rackspace Technology (NASDAQ: RXT) is a heavily leveraged hybrid multicloud and managed services provider that is trying to reposition itself as an “AI-first” infrastructure and operations partner. After years of revenue decline, recurring losses and a difficult competitive landscape, the February 18, 2026 strategic partnership with Palantir has triggered a violent rerating of the stock and reopened the turnaround debate.
SpaceX launches satellites, Starlink keeps front lines online, X shapes the information space, and xAI’s Grok is being wired into military networks. A new Pentagon contest for voice-controlled autonomous drones shows how these pieces are starting to lock together into a single, highly political war-tech stack.
Recursion, Tempus AI and Exscientia are three “pure-play” examples of what it really means to put artificial intelligence at the centre of a biotech business model. Used properly, AI can compress timelines, surface non-obvious biology and make better use of messy data. Used badly, it can amplify errors, burn cash faster and create stories the clinic will never confirm.
The logic of this basket is to follow the AI value chain from data to models to deployment, without touching the mega-cap hyperscalers. Each ticker is plugged into a different piece of that chain: some provide platforms, some sell services, some sell hardware that enables AI workloads.
The core idea is blunt but useful: markets are paying up for the “picks and shovels” of AI – chips, foundry, high-bandwidth memory, data centres, cybersecurity, observability – and they are putting a heavy discount on businesses whose main value proposition can be replicated by AI agents, plug-ins and low-cost automation. It is not just about which tickers are in fashion; it is about which roles and workflows are becoming cheaper to automate.
Small-cap geospatial intelligence pure play with a fast-growing Gen-3 satellite constellation, recurring U.S. and international defense contracts and strong leverage to the new U.S. space and missile-defense cycle.
BigBear.ai started as a data-fusion and analytics contractor, grew via acquisitions and went public through the GigCapital4 SPAC at the end of 2021. The company’s value proposition is to make sense of messy data streams and provide actionable recommendations for operators in defense, intelligence, border security, logistics and other complex environments. To do that, it combines traditional analytics, AI/ML models, simulation, and mission-planning tools into a single decision-intelligence laye
2026 opens in a geopolitical environment that looks like a “Cold War 2.0” – but running at a much higher speed and across more domains than anything we have seen before. Classic threat pillars – nuclear deterrence, air and sea control, conventional superiority – are still there, but they have been reshaped by disruptive technologies that make them far more lethal and unpredictable.
