The semiconductor landscape is proceeding towards a highly transformative state. A oneGbps speed is something not unheard of for regular internet users these days. The emergence of Ultra high-speed broadband technologies such as G.fast, fiber technology and the Passive Optical Network (PON) have all facilitated this. We also increasingly see broadband wireless to handheld devices such as a Smartphone or Tablet using newer versions of 4G LTE. However, there is a new push to go to 5G Wireless, which will bring bandwidths of up to oneGbps to users on handheld devices in the future.
The first generation of G.fast will consume 106MHz of bandwidth and provide approximately gigabit rates at about 100 meters, and will see wide deployment in North America and the Western Europe. In countries like Japan and Korea, they opt for a different flavor of copper technology better suited for their deployment scenario in multi-dwelling units (MDUs). This in cohesion with the already widespread deployment of the 30MHz version of VDSL2, which provides a 100Mbps symmetric service to residential customers today on the same copper network. However there will be life beyond the first generation of G.fast. The International Telecommunications Union is working on the development of the second-generation G.fast standard, which will consume up to 212MHz and will provide nearly twoGbps at about 50 meters. So copper may continue to live for years to come! Devices such as smart meters or smart refrigerators that fall into the IoT category and sensors such as gyroscopes or accelerometers that are being embedded into smartphones and tablets are the next set of trends that we see as the industry evolves.

IoT- Driving the Technology Innovation
Today, the technology innovation is being driven more by the needs of consumers, including what is needed in a so called "connected home." There are a couple of things that are very much needed now. A couple of examples of these are:
1. Guaranteed ultra-high-speed (hundreds of megabits/sec or even a gigabit/sec) broadband connections coming into homes.
2. Distribution of high-speed bandwidth within the home to enable a simultaneous multi-screen HD, 4K or 8K experience.
3. An IoT infrastructure including ultra-low-power sensor silicon and centralized control of these form a gateway for home automation. Data collection and diagnostics should also be supported by this IoT network at home for analytics to be performed in the cloud to facilitate the service providers' ability to better serve the consumer. The ultra-low-power sensors should be battery-operated so that they can be placed anywhere in the house as appropriate. These should not need a change of battery for at least a year, if not for five years.
On the enterprise/infrastructure side, one of the big unsolved problems is backhaul for Pico cells for ubiquitous high-bandwidth 4G-LTE coverage. LTE promises ultra-high-speed wireless connections. However the availability of high speed can only be guaranteed if a Pico cell is nearby. In order to provide ubiquitous coverage, these must be placed densely, e.g., every 250m or so. However, these cells must be fed by high-speed links which do not exist today! This problem will potentially be carried over to 5G wireless as well. Ubiquitous availability of gigabit/sec service to a few wireless users or tens of megabits/sec to thousands of wireless broadband users will never happen without the availability of proper backhaul; one way to achieve this would be by feeding the Pico cells with fiber or with G.fast or with bonded G.fast. So in some sense the success of wireless depends on wireline ultra-broadband.

Higher Capital, Hard to Come By
The biggest challenge anyone faces is the amount of capital needed to develop a complex IC chip today. Given that just a single 16nm tape-out of any reasonably complex device costs several million dollars, it is increasingly getting difficult for startups or even smaller semiconductor companies to be able to compete with the larger companies.Often referred to as "Cost of ASIC" problem that explains how expensive it has become to develop custom complex ASICs today using deep submicron technology, the industry has yet to come up with a solution for this. A major concern of the industry is the worry about ROI. However, there may be hope, as some folks may be thinking of alternatives to full custom ASICs to deal with the enormous cost of developing these chips today. One of the startups, Wave Semiconductor of Campbell, CA is focusing solely on solving the "High Cost ASIC" problem and has taken on this problem head on.
On a different note, not every chip may have to use the latest process geometry, and older fab processes/ geometries may just be fine for those applications at hand. Also many analog chips or MEMs do not use expensive processes. IoT devices or sensors largely fall in this category, and hence we may see a real growth opportunity for semiconductor startups in this arena going forward.
There are a couple of challenges that have been very prominent lately and which many semiconductor peers would agree to:
(a) The products we develop require a lot of R&D. Developing and abiding by complex technologies is quite a challenge, and there is lack of monetary encouragement for us. The challenge is to get paid "well" or even "fairly" for the value of the complex technology and products we bring to the table and still make money and be profitable.
(b) How to attract investments; inherently we have long cycles for return on investment, unlike in the software industry.
(c) A vision for success; reinventing the company to diversify into new or adjacent markets by leveraging our core competency and augmenting it appropriately with new talent.

Encouraging Multiple Ways
I always encourage having a two-way approach for solving a specific challenging task. Through that we can compare different solutions and then pick the one that best serves our needs.I value intellectual property. So I encourage people to develop new ideas, to emphasize and the strategic areas of interest for the company as far as intellectual property is concerned which, if found to be novel would lead to the filing of provisional patents. We also recognize our folks for filing patents, as well as for when the patent is granted in various ways. This motivates people to think outside the box, and to invent and generate intellectual property that is valuable for us.