BigBear.ai Secures $225M in Contracts, Cash at $409.8M

Key Takeaways

• BigBear.ai won over $225 million in contracts with the FBI, NASA, Coast Guard, DOE, and VA
• Innodata reported 58% YoY revenue growth to $92.1 million but trades at a premium P/E of 35.47X
• Micron guides Q1 fiscal 2027 revenue at $61.5 billion versus NVIDIA's $108 billion forecast
• Proofpoint to acquire Varonis for $6.3 billion to build an AI-security platform
• Grok Voice 3.0 ranks first on Artificial Analysis with support for 20 languages

BigBear.ai trades below industry as AI demand and contract wins grow

BigBear.ai is trading below its industry average with a forward P/S ratio of 6.52. The company's stock has fallen 59.6% year to date. It secured over 20 new contracts in Q2 2026 and grew backlog to $269.6 million. The company also holds about $409.8 million in cash and investments.

Morgan Stanley beats Q3 earnings with strong wealth and AI focus

Morgan Stanley reported Q3 earnings of $3.03 per share, beating the estimate of $2.83. Revenue reached $20.34 billion, up about 12% year over year. Wealth Management revenue hit a record $9.1 billion with $232 billion in net new assets. CEO Ted Pick highlighted AI infrastructure investment as a major growth opportunity.

Innodata stock faces high valuation despite strong AI revenue growth

Innodata is trading at a forward P/E ratio of 35.47X, above its industry average of 23.19X. Q2 2026 revenue rose 58% year over year to a record $92.1 million. The company expects at least 40% revenue growth for 2026. Its stock has fallen 34.3% over the past year despite expanding AI services.

Micron vs NVIDIA: Which AI stock is the better buy after earnings

Micron reported fiscal Q4 2026 revenue of $54.23 billion and a non-GAAP gross margin of 87%. It guides Q1 fiscal 2027 revenue at $61.5 billion. NVIDIA posted Q2 fiscal 2027 revenue of $96 billion and expects $108 billion next quarter. Micron has lower leverage and a lower forward P/E ratio of 5.18 versus NVIDIA's 23.93.

BigBear.ai wins over $225M in contracts across defense and intelligence agencies

BigBear.ai secured more than $225 million in contract awards since July 2026. The wins include over 25 new generative AI contracts with the FBI, NASA, U.S. Coast Guard, Department of Energy and Department of Veterans Affairs. The company also retained CBP work through two recompetes. Total backlog stood at about $270 million as of June 30, 2026.

Proofpoint to buy Varonis for $6.3 billion

Proofpoint and Varonis announced a deal where Proofpoint will acquire Varonis for $6.3 billion. The purchase is expected to close in the fourth quarter of 2022. The combined company aims to create a complete security platform against AI-accelerated threats and data risks. The deal is expected to be profitable for Proofpoint in the first year with cost synergies of about $50 million.

Zero trust and accountability needed for AI security

An article argues that AI must be treated as an international security problem requiring new rules. It says enterprises should use zero-trust principles to manage AI risks proactively. The piece highlights three critical risks from AI misuse that creators are struggling to control. It calls for clear accountability in AI cybersecurity frameworks.

China urges AI cooperation with Apple and Tencent leaders

Chinese Vice Premier Ding Xuexiang met with Apple CEO Tim Cook, Tencent founder Pony Ma, and Baidu founder Robin Li in Beijing. He called for stronger international cooperation on AI development and governance. Ding urged support for creating the World Artificial Intelligence Cooperation Organization. Nvidia CEO Jensen Huang and AMD CEO Lisa Su were also recently added to the advisory board.

California leads the way on AI regulation

California has passed new laws to regulate the use of artificial intelligence. Governor Gavin Newsom signed the legislation that was passed by the state legislature in August. The laws aim to ensure AI systems are transparent, accountable, and fair. They also include provisions to protect privacy and prevent misuse of AI.

Grok Voice 3.0 ranks first on Artificial Analysis

SpaceXAI released Grok Voice 3.0, a speech-to-speech model that ranks first on the Artificial Analysis Speech-to-Speech Index with a score of 86.9. The model listens, reasons, calls tools, and speaks. Developers can use it via WebSocket, WebRTC, and SIP with a price of $0.08 per minute. It features 28 built-in voices and supports 20 languages.

Celebrity AI deepfakes spread widely on social media

AI-generated photos and videos of celebrities are getting millions of views on platforms like Instagram. Fake accounts have used images of Mia Goth, Scarlett Johansson, and Taylor Swift, sometimes in sexualized or misleading contexts. Google search results even spread AI-made claims about Goth. Meta removed some accounts after The Verge asked about the problem, but AI deepfake content keeps growing online.

AI stock sell-off worries investors but boom continues

Tech stocks fell for a second day, but experts say this AI market scare may be temporary. Anthropic reported strong revenue numbers before its IPO, and TSMC posted record profits. Investors grew concerned about Core Weave debt notes, but CEO statements and support from Jensen Huang helped ease fears. Analyst Patrick Moorhead says this feels like past scares such as the Deepseek and Broadcom crashes.

Quantum computer models struggle with credit card fraud detection

A new study tested Quantum Kernel Support Vector Machines for detecting credit card fraud. Classical models reached a top PR-AUC of 0.983 with full data, but quantum models scored only 0.62 on real hardware due to gate errors and noise. The research found that scaling and feature limits are major bottlenecks for near-term quantum machine learning in finance. Synthetic oversampling helped improve results but could not close the gap.

New tool diagnoses problems in multimodal AI benchmarks

Multimodal benchmarks test AI models that use text, images, and other data types together. A new framework called SRCoD checks whether benchmark items truly need multiple modalities or can be solved with shortcuts. The method produces item-level diagnostic profiles and provides clear explanations for its findings. Experiments showed SRCoD performs better than existing diagnosis methods.

Sources

• BigBear.ai Trades at a Discount to Industry: Buy the Stock or Hold?
• Morgan Stanley Q3 Earnings Call Highlights Wealth and AI Strategy
• Innodata Stock's Premium Valuation: Can AI Growth Justify the Price?
• Micron vs. NVIDIA: Which AI Stock Is the Better Buy After Earnings?
• BigBear.ai Secures More Than $225M in Contract Awards Across Homeland Security, Defense, Intelligence and Federal Civilian Agencies
• Proofpoint to Acquire Varonis in a $6.3 Billion Transaction
• Zero trust and clear accountability are the new baseline for AI cybersecurity
• China's vice premier urges AI cooperation in meeting with Apple, Tencent chiefs | MLex | Specialist news and analysis on legal risk and regulation
• California takes the lead on regulating AI
• Meet Grok Voice 3.0: SpaceXAI Voice Agent Model Ranks #1 on Artificial Analysis
• Nobody is safe from celebrity AI deepfakes
• This AI rout is just another scare for investors
• Benchmarking Quantum Kernel SVMs for Credit Card Fraud Detection: Performance Realities Under NISQ Constraints
• Are Multimodal Benchmarks Really Useful? Item-Level Multimodal Benchmark Diagnosis via Structure-Response Co-Calibration