what to buy at this point
Recently, there’s been a fascinating phenomenon in the AI sector: discussions about “AI” are heating up, but capital isn’t flowing evenly to every AI company. Large tech giants, hot AI chip stocks, and a handful of star companies have grabbed most of the attention. Meanwhile, companies sitting at different stages of the AI supply chain are getting far less notice.
Yet there are still AI-related stocks with clear undervaluation potential. “Undervalued” here doesn’t mean “cheap,” nor does it guarantee an imminent rise — it simply means a company’s AI business, growth prospects, or asset value may not be fully reflected in its current market price.
I focus on three key factors: whether AI can actually translate into real business, whether the company’s cash and assets can sustain growth, and whether there are clear catalysts on the horizon.
1. $GOOG — The AI giant that’s easiest to overlook
Most people’s first reaction to Google is still search, YouTube, and ads. But I think the more interesting story now is that Google is quietly transforming from an internet company into a complete AI infrastructure powerhouse.
In Q1 2026, Alphabet reported $109.9 billion in revenue, up 22% year-over-year, with Google Cloud surging 63%. More importantly, Google Cloud is now supplying multi-gigawatt-scale TPU computing hardware to select customers. Its own AI models (Gemini), Cloud platform, and TPU chips are forming a tight, self-contained ecosystem.
Of course, $GOOG has challenges. The company raised its 2026 capex guidance to $195–205 billion, and the DeepMind leadership changes have raised questions about execution. The market worries Google might lose the AI race.
But that’s exactly why I find it compelling: Google already owns search traffic, Android, YouTube, Cloud, proprietary chips, data centers, and Gemini — a full-stack AI ecosystem.
As of today, $GOOG is still down roughly 15% from its May 2026 high of $404.47.
2. $BBAI — Unsexy, but AI is finally moving into its real strength
$BBAI is probably the most overlooked name on this list.
BigBear.ai focuses on defense, security, intelligence, trade, and travel — not consumer-facing AI. Its real value lies in whether AI can solve real problems for governments and large institutions.
In Q1 2026, revenue was $34.4 million (basically flat YoY), but gross margin jumped from 21.3% to 34.0%. Backlog grew 14% to $281.9 million, including a $53 million single-source secret project contract. The company ended the quarter with ~$43.15 million in cash and investments and maintained its full-year 2026 revenue guide of $135–165 million.
It’s still losing money and must be watched for past financing, debt, and dilution issues.
Yet the market often fixates on “how much money it’s making right now” and ignores the growing backlog, government AI demand, and high-margin business.
I’m not looking for $BBAI to become the next $PLTR. I’m watching whether it can turn project-based revenue into more stable, long-term contracts. If it does, the relatively small current market cap could leave plenty of room for re-rating.
3. $RXRX — The AI + pharma stock worth watching for the long term
If $GOOG represents AI infrastructure, then $RXRX represents the completely different space of AI + biopharma.
Recursion’s real edge isn’t just “using AI to help scientists” — it’s building an end-to-end AI-driven drug-discovery platform that combines massive experimental data with computational power to find new targets and candidates.
In Q1 2026, the company said multiple internal and partnered programs continued advancing: REC-1245 entered clinical stage, and REC-4881’s Phase II data showed strong efficacy signals. The company expects 2026 operating cash burn to stay under $390 million and believes its cash reserves can fund operations into early 2028 without new financing.
Equally important, its partnership with Sanofi has already produced real milestone payments. As of early 2026, Sanofi-related projects had generated $134 million in cumulative payments, with several AI-driven small-molecule programs in active development.
Of course, $RXRX is still high-risk. AI drug discovery must go through clinical trials, regulatory approval, and commercialization — a long, uncertain road. AI can speed things up, but it can’t remove the inherent uncertainty of biotech.
That’s precisely why I view $RXRX as a long-dated “option on how AI will change drug R&D.” If one or two core assets succeed, the market could re-rate the entire platform.
4. $SOUN — AI is moving from chatbots to the real world
For the past few years, when people talked about AI, they mostly meant ChatGPT, Gemini, or GPUs.
But I’m increasingly convinced the next really interesting question is: how will AI enter cars, restaurants, customer service, devices, and everyday life?
That’s exactly why $SOUN is worth watching.
SoundHound’s core focus is voice and conversational AI. The company is expanding into agentic AI and enterprise applications. At the end of Q1 2026 it had ~$216 million in cash, zero debt, and guided full-year revenue of $225–260 million.
In July 2026, Gartner named SoundHound a Leader in the Magic Quadrant for Conversational AI Platforms — proof it’s already carving out a solid position in enterprise voice AI.
Valuation isn’t cheap by traditional standards — at ~$3 billion market cap as of August, it’s no micro-cap.
But it has a powerful moat: it’s not competing with OpenAI on who has the smarter chatbot. It’s hunting for the commercial entry points where AI actually touches the physical world. If voice in cars, restaurants, customer service, or AI agents ever scale, $SOUN is positioned to capture a big slice of that growth.
5. $MAAS — Highest risk, but also the one most likely to be re-rated
Compared with the other four, $MAAS is a completely different investment thesis.
In March 2026, MAAS completed the acquisition of Huazhi Future assets and shifted strategy toward building a full AI chain: compute infrastructure → algorithms → intelligent hardware → real-world applications.
In April it launched the Stars Distributed Intelligent Computing Center project (up to 50 billion RMB investment). In July its subsidiary signed an MOU with Kazakhtelecom (KT-Telecom) to build a large AI data center in Kazakhstan.
From a narrative standpoint, this perfectly matches today’s AI infrastructure trend: as models get stronger, the need for compute, data centers, power, and localized deployment grows.
But I must highlight the risk: $MAAS faces higher hurdles than $GOOG, $BBAI, or $RXRX. It paid $26 million cash plus issued a large number of shares to acquire Huazhi Group, and in July it announced the sale of its 49% stake in Laixi Intelligent Technology for $17 million cash.
In short, $MAAS must prove that its AI compute and data-center projects can actually generate revenue, cash flow, and profit — not just good stories.
By August 2026 the stock had already risen nearly 200%, so it’s no longer a forgotten micro-cap, but it still offers enormous re-rating potential compared with big AI names — provided the company keeps executing.
Conclusion
If I put all five together, I wouldn’t simply ask “which one is best.”
I’d view them as five distinct AI investment philosophies:
- $GOOG: AI infrastructure + models + Cloud + massive cash flow
- $BBAI: AI + defense/security + government contracts
- $RXRX: AI + drug discovery + long-term tech platform
- $SOUN: AI agents + voice + real-world applications
- $MAAS: AI compute + data centers + China & overseas infrastructure
They carry completely different risk profiles.
For relatively mature AI business models, I’d lean toward $GOOG. For higher-beta small- and mid-cap AI names, I’d research $BBAI, $RXRX, and $SOUN more closely. $MAAS is the highest-risk, highest-re-rating-potential story of the bunch.
The most important point: “undervalued” does not mean “cheap stock.”
True undervaluation is when the market price sits well below the cash flows, technology value, and industry positioning the company is capable of creating in the future.
That’s why I believe these five names deserve continued observation.
