Piyumi Kapugeekiyana

application/vnd.openxmlformats-officedocument.wordprocessingml.document

From simple chats to complex conversations ChatGPT and the future of OpenAI

Chatgpt

From simple chats to complex conversations: ChatGPT and the future of OpenAI

<div>
 <ul>
 <li>
 
 Classical AI&mdash;particularly machine learning (ML)&mdash;is the most widely used technology across almost all focus areas
 
 </li>
 <li>
 
 The application of generative AI (GenAI) has accelerated, making it the technology with the second-highest adoption rate
 
 </li>
 <li>
 
 R&amp;D saw the most technology adoption, particularly in the use cases of drug discovery, genetic testing, drug development, and clinical trials
 
 </li>
 </ul>
 
 
</div>

<div>
 <ul>
 <li>
 
 Classical AI&mdash;particularly machine learning (ML)&mdash;was the most widely used technology across almost all focus areas. ML alone was present in over 40% of activity, with classical AI as a whole (including computer vision and natural language processing [NLP]) accounting for around half of our database. Big Techs, particularly, Alphabet and Microsoft, have been active as well. We believe that recent moves by the FDA and the European Medicines Agency (EMA) to accommodate drug applications that incorporate AI and ML components support increased adoption and investment in AI-based technologies.
 
 </li>
 <li>
 
 The application of generative AI (GenAI) has accelerated, making it the technology with the second-highest adoption rate at nearly 25%. It is most widely used in R&amp;D, particularly in drug discovery and development.
 
 </li>
 <li>
 
 In terms of focus area, R&amp;D saw the most technology adoption, particularly in the use cases of drug discovery, genetic testing, drug development, and clinical trials. 
 
 </li>
 </ul>
 
 
</div>

State of tech in Healthcare: How AI is transforming the entire value chain

Property owners and real estate companies are able to handle larger portfolios while offering faster and cheaper services.

Residential PropTech

<div>
 
 A deep dive into Notarize's financials, product, competitive environment, and market opportunities...
 
</div>

<div>
 
 
 <ul>
 <li>
 
 Notarize generated USD 80 million in revenue in 2023, growing 33% YoY. During the Covid-19 pandemic, revenue surged 600% (March 2020&ndash;early 2021), leading to a USD 130 million Series D at a USD 760 million valuation.&nbsp;
 
 </li>
 </ul>
 
 
 <ul>
 <li>
 
 Notarize is a digital platform for 24/7 remote online notarization, connecting users with licensed notaries via video calls. It operates on a transaction-based model, charging USD 10 per notarization (USD 3 for extra seals) and USD 25 for real estate closings. Notaries set their own rates, keeping the difference from Notarize's fees, incentivizing high-quality notary participation.
 
 </li>
 </ul>
 
 
 <ul>
 <li>
 
 Notarize faces competition from traditional notary services (UPS, Fedex), remote online notarization (RON) platforms (DocVerify, NotaryCam), and broader digital transaction management solutions (Docusign, Adobe Sign).
 
 </li>
 </ul>
 
 
 <ul>
 <li>
 
 Notarize faces three risks: 1) regulatory shifts &ndash; state-by-state regulations could change, with some states potentially reversing remote notarization acceptance, shrinking the market; 2) commoditization &ndash; larger players like DocuSign or Adobe could enter the market, bundling notarization services at lower prices, pressuring margins and reducing competitive advantage; and 3) trust issues &ndash; a high-profile fraud case could damage trust in remote notarization, leading to regulatory scrutiny and hesitancy from enterprise customers.
 
 </li>
 </ul>
 
 
</div>

Notarize: The online remote notarization service

<div>
 
 A deep dive into Ramp revenue, product, competitive environment, risks and market opportunities...
 
</div>

<div>
 <ul>
 <li>
 
 Sacra estimates Ramp reached USD 295 million in net annualized revenue in 2023, up 136% from USD 125 million in 2022. Total payments volume (TPV) grew 209%, from USD 10 billion in 2022 to USD 30 billion in 2023, driven by card volume growth (USD 18 billion from USD 8 billion) and bill pay expansion (USD 12 billion from USD 2.5 billion). Ramp primarily earns revenue from interchange fees with a gross take rate of 2.80%.
 
 </li>
 <li>
 
 Key platform features include 1) automatic expenses: text-based receipt uploads for automatic categorization and record-keeping, 2) policy enforcement: rules for spending by card, vendor, category, or department, automatically enforced, and 3) invoice digitization: OCR technology to quickly process and reconcile invoice details like vendor, amount, and items.
 
 </li>
 <li>
 
 Ramp faces competition from banking-as-a-service firms (Marqeta, Unit) and Brex (a business expense management platform). Unlike Amex's weeks-long process, Brex enables startups to get instant corporate card approvals via Plaid. Also, unlike Amex's USD 300&ndash;600 annual fees, Brex offers free card and startup perks like AWS credits and monetizes through transaction volume.
 
 </li>
 </ul>
 
 
</div>

Ramp: The corporate card and expense management company

Retail investors demanding sophisticated passive investment strategies are looking for an edge.

Retail Trading Infrastructure

<div>
 
 A deep dive into Betterment's financials, product, competitive environment, risks and market opportunities...
 
</div>

<div>
 
 
 <ul>
 <li>
 
 Betterment's revenue grew 69% to USD 153 million in 2023, up from USD 91 million in 2022, driven by increased AUM from deposits into its robo-advisor and Betterment Reserve high-yield cash account. Rising interest rates from 2022 to 2023 helped neobanks like Betterment earn more on deposits, enabling customer acquisition by passing some of these earnings on.
 
 </li>
 </ul>
 
 
 <ul>
 <li>
 
 Betterment generates revenue through five streams: 1) portfolio management fees (largest revenue component), 2) interchange fees, 3) partner banks, 4) advice packages, and 5) employers.
 
 </li>
 </ul>
 
 
 <ul>
 <li>
 
 Betterment faces competition from major players like Charles Schwab and Vanguard, which dominate robo-advisors with lower fees and vast networks. Competitors include Wealthfront and Personal Capital, FinTechs like Acorns, Chime, and Robinhood, and niche robo-advisors like Ellevest (for women) and Blooom (IRA/401k specialists).
 
 </li>
 </ul>
 
 
 <ul>
 <li>
 
 Risks include 1) lower interest rates: Betterment's 69% revenue growth in 2023 was fueled by high interest rates &ndash; if rates decline, it will need new revenue streams or strategies to retain users long term and 2) a lack of differentiation &ndash; with competitors offering similar products at comparable fees, Betterment risks losing its edge as its core investment services become commoditized.
 
 </li>
 </ul>
 
 
</div>

Betterment: The investment app with automated investing, financial planning, and cash accounts

<div>
 
 A look at NVIDIAs Q3 FY2025 announcements and their implications for EDGE industries..
 
</div>

<div>
 <h1>
 Key takeaways
 </h1>
 <ul>
 <li>
 
 <a href="https://sp-edge.com/industry/178">
 GenAI Infrastructure
 </a>
 &nbsp;
 
 <ul>
 <li>
 
 Announced plans to scale the output of the next-generation Blackwell GPU, aiming to deliver increasing quantities each quarter. Notably, the company had already shipped 13,000 Blackwell samples to customers during the quarter.
 
 </li>
 <li>
 
 A surge in infrastructure demands has resulted in NVIDIA 
 <a href="https://www.theverge.com/2024/5/22/24162860/nvidia-ai-chip-every-year-blackwell-rubin">
 accelerating
 </a>
 its development timelines, committing to release new chips annually instead of the previous two-year cadence.
 
 </li>
 </ul>
 </li>
 <li>
 
 <a href="https://sp-edge.com/industry/133">
 Next-gen Mobile Networks
 </a>
 &nbsp;
 
 <ul>
 <li>
 
 Launched 
 <a href="https://blogs.nvidia.com/blog/ai-aerial-wireless-networks/">
 AI Aerial
 </a>
 during Q3, offering AI-driven radio access network (AI-RAN) solutions to meet growing 5G and 6G infrastructure demands.
 
 </li>
 </ul>
 </li>
 <li>
 
 <a href="https://sp-edge.com/industry/149">
 Digital Twin
 </a>
 &nbsp;
 
 <ul>
 <li>
 
 Announced partnerships with Foxconn, Toyota, and Reliance to utilize NVIDIA Omniverse to create digital replicas of factories, enabling more efficient workflows and capitalizing on strengths in graphics processing and real-time simulation.
 
 </li>
 </ul>
 </li>
 <li>
 
 <a href="https://sp-edge.com/industry/185">
 Humanoid Robotics&nbsp;
 </a>
 
 <ul>
 <li>
 
 Launched 
 <a href="https://blogs.nvidia.com/blog/robot-learning-humanoid-development/">
 AI simulation tools
 </a>
 to advance robot learning and humanoid development, complementing its GPUs' strengths in powering AI workloads, particularly in computer vision&mdash;a crucial technology that enables robots to &quot;see&quot; through camera scans and image-processing ML algorithms.
 
 </li>
 </ul>
 </li>
 <li>
 
 <a href="https://sp-edge.com/industry/24">
 Quantum Computing
 </a>
 
 <ul>
 <li>
 
 SPEEDA Edge&rsquo;s proprietary 
 <a href="https://sp-edge.com/companies/4353#activities">
 activities tracker
 </a>
 showed partnerships with 
 <a href="https://nvidianews.nvidia.com/news/nvidia-supercharges-google-quantum-processor-design-with-simulation-of-quantum-device-physics">
 Google
 </a>
 , 
 <a href="https://sp-edge.com/updates/36154">
 AlgorithmQ
 </a>
 , and 
 <a href="https://sp-edge.com/updates/34200">
 Equal1
 </a>
 to utilize CUDA-Q&mdash;a software toolkit for orchestrating large-scale quantum computing applications, underscoring NVIDIA&rsquo;s efforts in the space.
 
 </li>
 </ul>
 </li>
 <li>
 
 <a href="https://sp-edge.com/industry/106">
 Hospital Management
 </a>
 
 <ul>
 <li>
 
 NVIDIA ventured into the area via 
 <a href="https://sp-edge.com/updates/34200">
 partnerships
 </a>
 with 
 <a href="https://sp-edge.com/companies/2555">
 Oracle
 </a>
 and 
 <a href="https://sp-edge.com/companies/120687">
 AvaSure Holdings
 </a>
 . By leveraging its accelerated computing infrastructure, the company aided the development of an AI-powered virtual concierge solution to enhance patient care and streamline clinical workflows.&nbsp;&nbsp;
 
 </li>
 </ul>
 </li>
 </ul>
 
 
</div>

NVIDIA Q3 FY2025 recap

Providing AI-driven market intelligence and workflow automation to give firms an edge

Capital Markets Tech

<div>
 
 This trend discusses how regulatory technology (RegTech) is streamlining labor-intensive processes related to Anti-Money Laundering (AML) compliance. It highlights the increasing complexity and scope of AML regulations, which now extend to industries such as gambling, insurance, and real estate in addition to financial services and the role of RegTech in simplifying the KYC and KYB processes.
 
</div>

<div>
 <ul>
 <li>
 
 A
 ML compliance requires 
 Know-Your-Customer (KYC)
 and 
 Know-Your-Business (KYB)
 processes, which are typically manual, time-consuming, and labor-intensive.
 
 </li>
 <li>
 
 RegTech solutions
 automates key compliance tasks, such as regulatory monitoring, reporting, and risk management, with a primary focus on AML requirements.
 
 </li>
 <li>
 
 RegTech addresses 
 customer identification, due diligence, and ongoing monitoring
 through tools for identity management, risk assessment, transaction monitoring, and regulatory reporting.
 
 </li>
 <li>
 
 RegTech firms are acquiring other RegTech companies to expand their service portfolios and market presence, focusing on functional areas like i
 dentity management, risk management, and transaction monitoring.
 
 </li>
 <li>
 
 Gen AI generates 
 authentic-looking fake content
 , such as social media personas and legal documentation, that could evade detection in AML compliance systems.
 
 </li>
 </ul>
 
 
</div>

RegTech

Cryptocurrencies

<div>
 
 Deep dive into Ripple's product, business model, competition, market opportunities, and risks
 
</div>

<div>
 
 
 <ul>
 <li>
 
 Ripple's xCurrent enables real-time, tracked cross-border payments with pre-validation and exchange rate locking, reducing settlement times to seconds. Its expanded suite includes On-Demand Liquidity (formerly xRapid) for sourcing cross-border liquidity and xVia for streamlined global payments. Together, these form RippleNet, connecting financial institutions and payment providers for seamless global transactions.
 
 </li>
 </ul>
 
 
 <ul>
 <li>
 
 Ripple earns revenue from xCurrent software licensing, XRP sales, and interest fees from institutional lending. xCurrent charges a USD 10 million installation fee and annual fees of USD 100,000&ndash;500,000 based on usage, enabling real-time international payments at lower costs. Ripple's edge lies in blending traditional finance with blockchain technology.
 
 </li>
 </ul>
 
 
 <ul>
 <li>
 
 Ripple&rsquo;s competitors include traditional financial infrastructure firms (JPMorgan, Citigroup), enterprise blockchain solutions (IBM&rsquo;s Hyperledger Fabric, R3&rsquo;s Corda), and cryptocurrency payment networks (Stellar).
 
 </li>
 </ul>
 
 
 <ul>
 <li>
 
 Ripple faces key risks, including limited bank adoption of XRP, regulatory uncertainty over its classification as a security (highlighted by the SEC lawsuit), and competition from central bank digital currencies (CBDCs), which could reduce demand for XRP as a bridge currency.
 
 </li>
 </ul>
 
 
</div>

Ripple: The blockchain-based digital payment company

The toolkit to build machine learning applications at scale

Machine Learning Infrastructure

<div>
 
 A deep dive into H2O.ai's product, competitive environment, risks and market opportunities...
 
</div>

<div>
 
 
 <ul>
 <li>
 
 H2O.ai found product-market fit as an open-source machine learning platform for data scientists, gaining early traction with financial firms like Capital One and Wells Fargo. The core platform enables model building and deployment in R and Python, supporting algorithms like deep learning and random forests on large datasets with parallel processing.
 
 </li>
 </ul>
 
 
 <ul>
 <li>
 
 H2O.ai uses a hybrid open-source/enterprise SaaS model, offering a free ML platform while monetizing through premium subscriptions and the H2O.ai AI Cloud platform. Enterprise pricing starts from USD 6,900 annually; the AI Cloud offering requires a USD 20,000 minimum commitment.
 
 </li>
 </ul>
 
 
 <ul>
 <li>
 
 The company faces competition from major cloud providers (Microsoft, AWS, Google Cloud), specialized ML companies (DataRobot, Databricks), and open-source-focused companies (Anaconda, RStudio).
 
 </li>
 </ul>
 
 
 <ul>
 <li>
 
 Key risks faced include open-source commoditization: 1) balancing community goodwill and monetization, as enterprise tools like Driverless AI risk competition from open-source alternatives, 2) AI platform consolidation; cloud providers like Microsoft, Google, and Amazon integrating ML/AI directly could reduce demand for standalone platforms, and 3) talent retention; competition for elite AI talent, including Kaggle Grandmasters, threatens H2O.ai&rsquo;s innovation and technical edge.
 
 </li>
 </ul>
 
 
</div>

H2O.ai: The open-source AutoML platform

Quantum information processing is slowly becoming a reality, ushering in a new technological era.

Quantum Computing

<div>
 
 A deep dive into PsiQuantum's product, business model, competitive environment, and market opportunities...
 
</div>

<div>
 
 
 <ul>
 <li>
 
 PsiQuantum, founded in 2015 by quantum physicists and photonics experts, is developing a scalable, fault-tolerant quantum computer using silicon photonics. Its system aims to scale to over 1 million qubits, enabling applications in drug discovery, materials science, and finance.
 
 </li>
 </ul>
 
 
 <ul>
 <li>
 
 PsiQuantum prioritizes building utility-scale quantum computers with millions of qubits for error-corrected algorithms, leveraging established silicon manufacturing and proprietary optical switching technology.
 
 </li>
 </ul>
 
 
 <ul>
 <li>
 
 PsiQuantum competes in the quantum computing hardware market against tech giants (e.g., IBM, Google, and Microsoft), quantum-focused startups (e.g., IonQ, Rigetti, and Quantinuum), and photonic quantum rivals like Xanadu.
 
 </li>
 </ul>
 
 
 <ul>
 <li>
 
 PsiQuantum faces risks from its exclusive reliance on silicon photonics, making it vulnerable to technical barriers and costly pivots due to deep integration with GlobalFoundries. Its focus on building million-qubit systems bypasses incremental milestones, risking delays and allowing competitors to secure early market advantages. Additionally, scaling novel components like photon sources, detectors, and high-performance optical switches involves significant manufacturing challenges, with potential yield and performance issues that could impact costs and timelines.
 
 </li>
 </ul>
 
 
</div>

PsiQuantum: A pioneer in advancing quantum computing hardware development

<div>
 <ul>
 <li>
 
 57% of banks use AI in fraud detection, and 48% in risk assessment. Major initiatives include Wells Fargo's AI in trade finance and Abound's USD 1 billion raise for AI lending.
 
 </li>
 <li>
 
 GenAI is growing, 69% of wealth management firms using GenAI in customer service.
 
 </li>
 <li>
 
 Blockchain and AI-enhanced biometrics are active, with 60%&ndash;75% biometric adoption in financial services.
 
 </li>
 </ul>
 
 
</div>

<div>
 <h1>
 Key takeaways
 </h1>
 <ul>
 <li>
 
 Classical AI (predominantly machine learning [ML] and natural language processing [NLP]) has found its footing in backend operations, 
 with
 
 57% of banks having deployed it in fraud detection and 48% in risk assessment. Notable developments included 
 <a href="https://www.fintechfutures.com/2024/04/wells-fargo-to-integrate-ai-into-trade-finance-operations-with-tradesun/">
 Wells Fargo
 </a>
 and 
 <a href="https://www.bobsguide.com/lloyds-and-cleareye-ai-forge-ai-centric-trade-finance-alliance/">
 Lloyds'
 </a>
 AI integrations in trade finance and significant funding rounds like 
 <a href="https://sp-edge.com/updates/31021">
 Abound's USD 1 billion
 </a>
 raise for international AI lending platform expansion
 
 </li>
 <li>
 
 GenAI is gaining momentum across customer-facing operations:
 Major initiatives included Wells Fargo's 
 <a href="https://itrexgroup.com/blog/generative-ai-in-banking/">
 Fargo
 </a>
 assistant (handling 100 million annual interactions) and NatWest's
 <a href="https://www.natwestgroup.com/news-and-insights/news-room/press-releases/data-and-technology/2024/jun/natwest-launches-cora-plus-the-latest-generative-ai-upgrade-to-t.html">
 Cora+
 </a>
 having practical applications of GenAI in this space. Notably, Temenos' 
 <a href="https://www.temenos.com/news/2024/05/13/temenos-launches-the-first-responsible-generative-ai-solutions-for-core-banking/">
 launch
 </a>
 of &ldquo;Responsible Generative AI&rdquo; solutions for core banking operations aimed to address regulatory compliance and data security challenges in financial operations.
 
 </li>
 <li>
 
 Blockchain and AI-enhanced biometrics appear as active trends across payments:
 Blockchain is making substantial progress in backend and customer-facing payment applications, with Worldpay reporting a 50% reduction in payment processing time through its stablecoin system and 29% of retail investors using crypto for transactions (up from 23% in 2022). Biometric adoption in financial services has reached 60%&ndash;75%&mdash;among the highest across sectors&mdash;with solutions like Mastercard's
 <a href="https://www.mastercard.com/news/press/2024/april/mastercard-transforms-the-fight-against-scams-with-latest-ai-tech/">
 Scam Protect
 </a>
 combining identity verification and behavioral biometrics.&nbsp;
 
 </li>
 <li>
 
 AI-driven personalization supports retail investment strategies and back-office automation for wealth managers:
 Product updates such as Revolut's 
 <a href="https://www.revolut.com/en-FR/news/revolut_launches_robo_advisor_in_eea_to_automate_investing/#:~:text=Revolut%20has%20made%20investing%20more,Advisor%20service%20in%20the%20EEA.">
 robo-advisory service
 </a>
 s and EarnUp's 
 <a href="https://sp-edge.com/updates/32186">
 AI Advisor
 </a>
 are examples of hyper-personalized investment products. GenAI adoption in wealth management is particularly strong in 
 customer service (69% of firms) and investment research (50%),
 with BCG estimating potential time savings of up to 30% in areas like customer boarding.
 
 </li>
 <li>
 
 Open banking data sharing receives big regulatory boost:
 Relaxation of regulations surrounding open banking has enabled banks to seamlessly share data among themselves, resulting in the emergence of new payment models, such as account-to-account (A2A) and multi-currency payments. Notable developments included 
 <a href="https://sp-edge.com/companies/41928">
 HSBC
 </a>
 &rsquo;s launch of 
 <a href="https://fintechmagazine.com/articles/hsbc-set-to-launch-international-payments-app-zing">
 Zing
 </a>
 , a 
 <a href="https://investor.visa.com/news/news-details/2024/HSBC-Partnered-with-Visa-to-Develop-Zing-International-Money-App/default.aspx">
 multi-currency payments app
 </a>
 , in collaboration with 
 <a href="https://sp-edge.com/companies/13061">
 Visa
 </a>
 
 </li>
 <li>
 
 Asset tokenization regains traction for institutional wealth management:
 The value of tokenized assets reached USD 15 billion as of September 2024 (excluding stablecoins). Major financial institutions, including BlackRock, Fidelity, and JPMorgan,&nbsp; actively implemented blockchain technology for tokenization, players such as State Street 
 <a href="https://www.paymentsjournal.com/state-street-to-launch-tokenization-platform-for-institutional-investors/">
 introduced tokenization
 </a>
 platforms, and partnerships between companies like AlphaPoint, Polymesh, 
 <a href="https://archax.com/insights/assetera-and-archax-announce-first-pan-european-strategic-partnership">
 Archax, and Assetera
 </a>
 aimed to enhance and facilitate asset tokenization.
 
 </li>
 <li>
 
 <a href="https://sp-edge.com/industry/24">
 Quantum computing
 </a>
 also showed promising use cases,
 particularly in portfolio optimization, simulations, and credit analysis. Key players such as 
 <a href="https://aws.amazon.com/blogs/quantum-computing/goldman-sachs-and-aws-examine-efficient-ways-to-load-data-into-quantum-computers/">
 Goldman Sachs
 </a>
 , 
 <a href="https://www.jpmorgan.com/technology/applied-research">
 JPMorgan
 </a>
 , and 
 <a href="https://www.hsbc.com/who-we-are/businesses-and-customers/hsbc-and-quantum">
 HSBC
 </a>
 were among the major institutions frontlining research into quantum computing in financial services. However, most real-world implementations are still in experimental phases, with 60% of regulators expecting a significant impact within as late as 
 <a href="https://www.theasianbanker.com/updates-and-articles/quantum-technology-can-revolutionise-finance,-but-create-new-problems">
 seven years
 </a>
 . Notably, 
 <a href="https://sp-edge.com/companies/29093">
 D-Wave Systems
 </a>
 , a maker of 
 <a href="https://www.dwavesys.com/solutions-and-products/financial-services/">
 quantum annealing systems
 </a>
 , is exploring applications in financial services, a hybrid approach that combines classical CPUs/GPUs with their quantum annealing hardware.
 
 </li>
 </ul>
 
 
</div>

EDGE Insights